High-Performance Computing

Research in high-performance computing (HPC) aims to design practical algorithms and software that run at the absolute limits of scale and speed. It is motivated by the incredible demands of “big and hairy” data-hungry computations, like modeling the earth’s atmosphere and climate, using machine learning methods to analyze every book and paper ever written, discovering new materials, understanding how biological cells work, simulating the dynamics of black holes, or designing city infrastructure to be more sustainable, to name just a few. To perform these computations requires new approaches for the efficient use of advanced computer systems, which might consist of millions of processor cores connected by high-speed networks and coupled with exabytes of storage. 


Specific HPC topic areas of interest within CSE include: 

  • Discrete and numerical parallel algorithms 
  • HPC applications in science and engineering 
  • Design of architecture-aware algorithms (e.g., GPU, FPGA, Arm) 
  • High-performance scientific software 
  • Performance analysis and engineering 
  • Post-Moore's Law computing 

HPC research at Georgia Tech is cross-cutting and multidisciplinary. CSE HPC faculty work closely with researchers across computing, social and natural sciences, and engineering domains. They lead interdisciplinary HPC research centers (see below) and contribute to HPC-driven domain specific research centers and institutes such as the Center for Relativistic Astrophysics (CRA) and the Institute for Materials (IMAT).


Related links:

Center for High Performance Computing (CHiPC)

Center for Research into Novel Computing Hierarchies (CRNCH)

Institute for Data Engineering and Science (IDEaS)

Interdisciplinary CSE Faculty specializing in high-performance computing research: